Ipsilateral increased oxygen extraction fraction (OEF) measured by PET is a powerful independent risk factor for subsequent stroke in patients with symptomatic carotid occlusion. The St. Louis Carotid Occlusion Study found that the relative risk conferred by increases OEF was 7.3 (95% Cl 1.6 - 33.4) for ipsilateral ischemic stroke corresponding to two year rates for ipsilateral stroke of 5.3% in 42 patients with normal OEF and 26.5% in 39 patients with increased OEF (p=.004). In patients with hemispheric symptoms within 120 days, the two year ipsilateral stroke rates were 12% in patients with normal OEF and 40% in patients with increased OEF. Previous PET studies have demonstrated that surgical anastomosis of the superficial temporal artery to the middle cerebral artery (STA-MCA) can restore OEF to normal. The Carotid Occlusion Surgery Study (COSS) will test the hypothesis that STA-MCA surgical anastomosis when added to best medical therapy can reduce by 40%, despite perioperative stroke and death, subsequent ipsilateral ischemic stroke (fatal and non-fatal) at two years in patients with recent (<120 days) symptomatic internal carotid artery occlusion and increased OEF. Clinically eligible patients with unilateral carotid artery occlusion demonstrated by any vascular imaging will be asked to consent to PET and treatment randomization if PET shows increased OEF. If necessary, consent for a supplemental arteriogram to confirm occlusion or demonstrate that extracranial and intracranial vessels will be sought as well. Those who agree and meet both arteriographic and PET criteria will be randomized 1:1 to undergo STA-MCA bypass surgery followed by medical management or medical management alone. All randomized patients will be followed at three month intervals for two years. The primary endpoint in the surgical group is the combination of the following: (1) 30 day post-operative (surgical group) or post-randomization (non-surgical group) occurrence of all stroke and death and (2) the occurrence of ipsilateral ischemic stroke within two years of randomization. Adjusting for anticipated 2 year mortality, 372 patients (186 in each group) will provide 90% power to detect the anticipated difference (40% vs. 24.2%). 91 have been randomized. Assuming 25-30% of PET scans will demonstrate increased OEF, this will require enrolling a total of 1400 clinically eligible subjects for PET. This proposal covers the second 5 years of this trial.